Abstract
We read with interest the article by Kaplan et al. comparing endobronchial intubation with the use of a bronchial blocker for pulmonary lobectomy in small children. 1 This retrospective study reported that single-lung isolation utilizing endobronchial intubations, compared with bronchial blockers, had shorter operative time, shorter anesthetic exposure, and less intraoperative blood loss. The study's conclusions stand in contrast to the findings of several larger prospective and retrospective studies. In addition, a lack of clarity on the number of providers and a lack of correction for multiple hypothesis testing represent significant limitations to their conclusions that endobronchial intubation is superior to using a bronchial blocker.
One randomized controlled trial in 60 patients comparing endobronchial intubation with use of a bronchial blocker for one-lung ventilation, in fact, revealed that endobronchial intubation led to longer placement times and required more repositioning. 2 In addition, a multicenter retrospective cohort study with 306 patients demonstrated that bronchial blockers were associated with fewer episodes of hypoxemia and less severe hypoxemia when compared with use of endobronchial intubation, although endobronchial intubation was preferred 2 to 1. 3
Generally, bronchial blockers are very effective at lung isolation, but can be technically challenging to position correctly in young children, and success placement is highly dependent on the expertise of the clinician. In Kaplan et al.'s study, there were only 32 subjects >9 years, and only 17 of those had bronchial blockers—this represents placing less than two bronchial blockers per year. It is difficult to imagine adequate proficiency and skill sustainment for any single clinician in placing a bronchial blocker when they are used rarely. Comparing this with Yan et al.'s study, the clinicians placed 30 bronchial blockers and 30 endobronchial intubations, and only 2 clinicians were involved to control for skill discrepancies. 2
It seems plausible that the relative infrequency in placing bronchial blockers might explain the findings of Kaplan et al.'s study and not an inherent flaw of the device. Moreover, it is not clear how many surgeons participated. This may represent a significant source of confounding, especially for blood loss endpoints, as accurate estimates of blood loss are difficult. Lastly and importantly, the authors do not describe any correction for multiple hypothesis testing, which may alter the author's results given the small sample size. For example, using a Bonferroni Correction for the 20 multiple hypotheses tested in the article (table 1) reduces the P value threshold for significance from P < .05 to P < .0025, which would negate the significance of all the study's findings.
Despite the limitations in this study, we agree with the authors that there needs to be more extensive randomized control studies across multiple centers to determine if one approach to lung isolation is superior. Unfortunately, small retrospective case series that draw conclusions without considering issues such as confounding, the impact of clinician competencies in rarely performed procedures, and a lack of corrections for multiple hypothesis testing can be misleading. They say in life, hindsight is 20/20, but in retrospective research, interpretations of the past can have significant limitations.